QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Data supplement All Versions of this Article: jpet.106.112268v1 320/1/266    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Koukouritaki, S. B. Articles by Hines, R. N. Search for Related Content PubMed PubMed Citation Articles by Koukouritaki, S. B. Articles by Hines, R. N.

METABOLISM, TRANSPORT, AND PHARMACOGENOMICS

Identification and Functional Analysis of Common Human Flavin-Containing Monooxygenase 3 Genetic Variants

Departments of Pediatrics (S.B.K., M.T.P., R.N.H.), Pharmacology and Toxicology (R.N.H.), and Biostatistics (N.M.P., T.W.), Medical College of Wisconsin, and Children's Research Institute, Children's Hospital and Health System (S.B.K., M.T.P., R.N.H.), Milwaukee, Wisconsin; and Department of Environmental and Molecular Toxicology, Linus Pauling Institute, Oregon State University, Corvallis, Oregon (M.C.H., L.K.S., S.K.K., J.E.V., D.E.W.)

Received August 9, 2006; accepted October 17, 2006.

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
Flavin-containing monooxygenases (FMOs) are important for the disposition of many therapeutics, environmental toxicants, and nutrients. FMO3, the major adult hepatic FMO enzyme, exhibits significant interindividual variation. Eighteen FMO3 single-nucleotide polymorphism (SNP) frequencies were determined in 202 Hispanics (Mexican descent), 201 African Americans, and 200 non-Latino whites. Using expressed recombinant enzyme with methimazole, trimethylamine, sulindac, and ethylenethiourea, the novel structural variants FMO3 E24D and K416N were shown to cause modest changes in catalytic efficiency, whereas a third novel variant, FMO3 N61K, was essentially devoid of activity. The latter variant was present at an allelic frequency of 5.2% in non-Latino whites and 3.5% in African Americans, but it was absent in Hispanics. Inferring haplotypes using PHASE, version 2.1, the greatest haplotype diversity was observed in African Americans followed by non-Latino whites and Hispanics. Haplotype 2A and 2B, consisting of a hypermorphic promoter SNP cluster (-2650C>G, -2543T>A, and -2177G>C) in linkage with synonymous structural variants was inferred at a frequency of 27% in the Hispanic population, but only 5% in non-Latino whites and African Americans. This same promoter SNP cluster in linkage with one or more hypomorphic structural variant also was inferred in multiple haplotypes at a total frequency of 5.6% in the African-American study group but less than 1% in the other two groups. The sum frequencies of the hypomorphic haplotypes H3 [15,167G>A (E158K)], H5B [-2650C>G, 15,167G>A (E158K), 21,375C>T (N285N), 21,443A>G (E308G)], and H6 [15,167G>A (E158K), 21,375C>T (N285N)] was 28% in Hispanics, 23% in non-Latino whites, and 24% in African Americans.

The individual FMO enzymes exhibit broad but distinct substrate specificities (Krueger and Williams, 2005) as well as species-, sex-, tissue-, and age-dependent differences in expression patterns (for review, see Hines, 2006). In the human, FMO3 is the predominant adult hepatic enzyme with a specific content of 60 ± 43 pmol/mg microsomal protein (Overby et al., 1997), comparable with the most abundant adult liver cytochrome P450 enzymes, i.e., CYP3A4 and CYP2C9 (Shimada et al., 1994). Also similar to the cytochromes P450, 10- to 20-fold interindividual differences in FMO expression have been described that may contribute to differences in toxicant susceptibility and/or therapeutic efficacy (Overby et al., 1997; Yeung et al., 2000; Koukouritaki et al., 2002; Hisamuddin et al., 2005). However, unlike the cytochromes P450, the contribution of FMO induction to interindividual differences in expression is at best, controversial. Hukkanen et al. (2005) described a significant increase in FMO3-dependent metabolism during pregnancy, whereas Zhang et al. (1996) reported an exacerbation of trimethylaminuria symptoms with menstruation, suggestive of FMO3 regulation by sex hormones. Yet, no gender differences were observed during postnatal FMO3 developmental expression (Koukouritaki et al., 2002). Most recently, Tijet et al. (2006) reported an 80-fold induction of FMO3 mRNA levels in male mice that was dependent on both the presence of the aryl hydrocarbon receptor and treatment with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Given that a strong sexual dimorphism exists in the mouse wherein FMO3 expression is suppressed in the adult male but not in the adult female (Cherrington et al., 1998), one must wonder whether the observed FMO3 induction was related to an effect on testosterone-dependent FMO3 suppression in this species.

The existence of polymorphisms affecting human FMO3-dependent metabolism has been well documented (for review, see Koukouritaki and Hines, 2005). Twenty-seven FMO3 allelic variants, including missense, nonsense, or deletion mutants, that result in a complete or near complete loss of functional activity are causative for trimethylaminuria or "fish-odor syndrome" (http://bsmsol2.biochem.ucl.ac.uk/Human_FMO3/). Because these alleles are rare in most populations, they cannot account for the substantial interindividual differences in FMO3-dependent metabolism. Several studies have documented reduced activity with the relatively common FMO3 E158K and E308G variants in vitro (Störmer et al., 2000) and in vivo (Park et al., 2002; Hisamuddin et al., 2005), although some substrate dependence is apparent. In some populations, these variants exhibit a high degree of linkage disequilibrium. When present on the same allele, the E158K and E308G exhibit an even more pronounced effect on FMO3 function (Park et al., 2002), even leading to mild or transient forms of trimethylaminuria (Zschocke et al., 1999). Several other functional FMO3 structural variants have been identified, but their frequency and as such, their relevance to population health is unknown (for review, see Koukouritaki and Hines, 2005).

In a recent single-nucleotide polymorphism (SNP) discovery study, we characterized 40 common FMO3 SNPs using DNA samples from the Coriell Polymorphism Discovery Resource (Camden, NJ) (Koukouritaki et al., 2005). Eleven of these polymorphisms were located within exon sequences, whereas seven novel promoter variants also were identified. The latter were used to infer seven promoter haplotypes that differ in frequency among the three different population groups studied: Hispanics of Mexican descent, African Americans, and non-Latino whites. Three of the inferred haplotypes significantly altered FMO3 promoter activity in vitro. To extend these findings, the objectives of the current study were to determine a more complete FMO3 haplotype structure in these same three population groups and to conduct functional analysis of structural variants predicted to impact FMO3 catalytic activity.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Materials. Platinum Taq DNA Polymerase High Fidelity, the BaculoDirect C-Term Expression kit, the pENTR/SD-TOPO Cloning kit, and Cellfectin were purchased from Invitrogen (Carlsbad, CA). The ExoSAP-IT mix, containing both exonuclease I and shrimp alkaline phosphatase, as well as shrimp alkaline phosphatase alone, was purchased from U.S. Biochemical Corp. (Cleveland, OH). CEQ SNP-Primer Extension and Dye Terminator Cycle Sequencing kits were obtained from Beckman Coulter, Inc. (Fullerton, CA). CleanSEQ magnetic beads were purchased from Agencourt (Beverly, MA). Custom oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA). Methimazole and ethylenethiourea were purchased from Lancaster Synthesis (Pelham, NH), whereas NADPH, FAD, sulindac, and sulindac sulfide were obtained from Sigma-Aldrich (St. Louis, MO). All other reagents were obtained from commercial sources at the purest grade available.

Patient Recruitment and DNA Isolation. Individuals representing various ethnic and/or racial groups were recruited to provide DNA samples as described previously (Koukouritaki et al., 2005). In all instances, ethnicity, race, or both were self-reported. Combining all DNA sources resulted in 202 samples from unrelated individuals of Hispanic (Mexican) descent, 201 samples from unrelated individuals of African-American descent, and 200 samples from unrelated individuals of non-Latino white (northern European) descent. Research protocols were approved by all Institutional Review Boards involved.

DNA Amplification. Template amplifications were performed in a final volume of 25 µl and contained 50 ng of genomic DNA, 2 mM MgSO₄, 0.2 mM each deoxyribonucleotide triphosphate, 0.2 µM each primer (see Supplemental Table S1), and 1 unit of Platinum Taq DNA Polymerase High Fidelity. All polymerase chain reactions had an initial denaturation step at 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and elongation at 68°C for 1 min. The final elongation step was performed at 68°C for 5 min.

Multiplexed Single Base Extension Assay. Genotype frequencies of FMO3 variants in different populations were determined using a multiplexed SBE assay as described previously (Koukouritaki et al., 2005). Sequences of the SBE primers used may be found in Supplemental Table S2. For a positive control, 20 fmol of linearized pRNH926 (reference sequence), pRNH927 (containing all queried SNPs), or an equal mixture of both were used as templates in the SBE reaction. The FMO3 sequence reported in National Center for Biotechnology Information accession number AL021026 [GenBank] was used as the reference for both the FMO3 gene, and, when translated in silico, the FMO3 enzyme.

Plasmids. An FMO3 variant 1 transcript, NM_001002294, position 46 to position 1736 was amplified from an adult human liver total RNA sample (Stratagene, La Jolla, CA) by reverse transcriptase-coupled polymerase chain reaction and cloned into the pCR2.1 vector to generate pRNH696. The fidelity of the cloned cDNA was verified by DNA sequence analysis on both strands. An FMO3 cDNA fragment, position +92 to +1698, was amplified using linearized pRNH696 as a template and 5'-GAT GAT TAG GTC AAC ACA AG-3' and 5'-CAC CAT GGG GAA GAA AGT G-3' as forward and reverse primers, respectively. The amplified product was cloned into pENTR/SD/D-TOPO, resulting in pRNH829. Site-directed mutagenesis was performed with the QuikChange site-directed mutagenesis kit (Stratagene) to introduce single-nucleotide variants in pRNH829, resulting in pRNH865 encoding FMO3 E24D, pRNH866 encoding FMO3 N61K, and pRNH904 encoding FMO3 K416N.

To produce a control plasmid for SBE genotyping that contains all 11 structural variants, a fragment representing FMO3 cDNA position 1 to 89 was amplified and cloned into pRNH829 immediately upstream of the existing cDNA fragment, generating pRNH926. Site-directed mutagenesis was performed using a QuikChange Multisite-Directed Mutagenesis kit (Stratagene) to introduce the desired single nucleotide changes into pRNH926, creating pRNH927. Nucleotide changes and the fidelity of nontargeted sequences were verified by DNA sequence analysis.

DNA Sequence Analysis. DNA sequencing was performed using a Dye Terminator Cycle Sequencing kit (Beckman Coulter, Inc.) according to the manufacturer's instructions. In brief, plasmid DNA templates were preheated at 96°C for 1 min followed by cooling to room temperature on the benchtop. Sequencing reactions were performed in a final volume of 20 µl and contained 50 fmol of plasmid DNA, 1.6 µM forward or reverse primer, and 8.0 µl of Dye Terminator Cycle Sequencing Quick Start Master Mix. All sequencing reactions included 50 cycles of denaturation at 96°C for 20 s, annealing at 50°C for 20 s, and elongation at 60°C for 4 min. Unincorporated primer and nucleotides were removed from the reactions using CleanSEQ magnetic beads (Agencourt) according to the manufacturer's instructions.

FMO3 Variant Expression. The FMO3 cDNA inserts from the pENTR clones were integrated with BaculoDirect linear DNA by LR-mediated Clonase II recombination. Sf9 cell transfection with the recombination reaction mixture was performed with Cellfectin in unsupplemented Grace's Insect Cell medium (Invitrogen). Negative selection was immediately begun with 0.1 mM ganciclovir. After amplification, viral stocks were used to infect Sf9 cells in SF-900 II SFM (Invitrogen) enriched with 10 µg/ml FAD. Cells were harvested 96 h postinfection. Microsomal fractions were prepared by differential centrifugation as described previously (Krueger et al., 2001) and resuspended in storage buffer (10 mM potassium phosphate, pH 7.6, 20% glycerol, 1 mM EDTA, and 0.4 mM phenylmethylsulfonyl fluoride). Protein concentrations were determined by the method of Bradford (1976). Flavin content, as a measure of FMO3-specific content, was determined as described previously (Henderson et al., 2004). In all instances, flavin measurements were corrected for a background of 0.083 nmol/mg microsomal protein based on the measurement of flavin-content in microsomal preparations from Sf9 cells infected with control vector.

Enzyme Assays. FMO3 catalyzed oxidation reactions were performed in 0.1 M Tris-HCl, pH 8.5, 1 mM EDTA, 0.1 mM NADPH, and with 40 to 500 µg/ml Sf9 microsomal protein. Trimethylamine (5–1200 µM) and ethylenethiourea (Lancaster) (10–100 µM) oxidation were assessed by following the oxidation of NADPH spectrophotometrically at 340 nm (^M = 6.22 · 10^–3 nmol/ml). The limit of detection for NADPH oxidation was 0.161 nmol/min/nmol FAD. Methimazole (5–1000 µM) oxidation was monitored at 412 nm following the method of Dixit and Roche (1984). The limit of detection for the methimazole assay was 0.035 nmol/min/nmol FAD. Sulindac (5–400 µM) oxidation was measured using high-performance liquid chromatography as described by Hamman et al. (2000) without differentiating between the S- and R-sulindac S-oxide. All reactions were incubated for 5 min, and the concentration of sulindac S-oxide was determined by linear regression using racemic sulindac S-oxide standards ranging between 0.1 and 10 nmol (r² = 0.999). Calculated rates were corrected for minimal rates of sulindac autooxidation observed in the presence of an equal concentration of non-FMO baculovirus-infected Sf9 microsomes.

Data Analysis. Identified SNPs were evaluated for deviation from Hardy-Weinberg equilibrium using a chi-square test. Allelic frequencies for individual sequence variants were compared using Fisher's exact test. A Bonferroni adjustment for the comparisons among the three groups was used, reducing the accepted value from 0.05 to 0.016. Individual haplotypes and their estimated population frequencies were inferred using PHASE, version 2.1 (Stephens et al., 2001; Stephens and Donnelly, 2003). Posterior estimates for unknown phase haplotypes were based on 150,000 iterations of the Gibbs sampler implemented in PHASE. In total, 5000 samples were discarded as a burn-in. Population recombination parameter estimates were based on 3000 samples using a thinning interval of 50 on the total iterations used to estimate subject haplotypes. The large thinning interval was to account for high autocorrelation among the sampled recombination rates. Geweke's convergence diagnostic (Geweke, 1992) was used to monitor for chain convergence. The Bayesian Output Analysis package for the R language was used to monitor the PHASE sample output both for convergence and autocorrelation among the posterior samples. Inferred mean haplotype frequencies were compared using a one-way ANOVA with a Holm-Sidak post hoc test (SigmaStat, version 3.1; Systat Software, Inc., Point Richmond, CA).

Analysis of molecular variance (AMOVA) was used to analyze the haplotype heterogeneity both within and among the three population samples. AMOVA was implemented using Arlequin, version 3.0 (Excoffier et al., 2005). The variance among populations is analogous to Wright's fixation index. Bootstrap confidence intervals were also computed based on 20,000 bootstrap replications. Pairwise values were also determined among the three populations to compare them individually, computing empirical P values in the same manner as the overall AMOVA. In addition, haplotype diversity (H) was evaluated using the estimated haplotype frequencies computed by PHASE as follows:

where n is the number of alleles in the population group, and p is the frequency of the ith haplotype of k total observed haplotypes for a given population. Asymptotic confidence intervals (95%) are presented for H using the sampling variance derived by Nei (1978).

Posterior estimates of population baseline recombination between adjacent SNPs () and multiplicative factor parameters (j) with 5, 10, and 25% posterior quantiles were calculated as described by Li and Stephens (2003).

Enzyme activity data at different substrate concentrations were fit to a single-site, nonlinear Michaelis-Menten model [goodness of fit (R²) > 0.99 in all instances], and kinetic parameters were determined using SigmaPlot, version 9.01 with Enzyme Kinetics Module, version 1.2 (Systat Software, Inc.). Log-transformed kinetic parameters for the different FMO3 variants were compared with the reference by one-way ANOVA with a Holm-Sidak post hoc test (SigmaStat, version 3.1; Systat Software, Inc.). Except as noted above, an of 0.05 was accepted as a significant difference between data sets. For the Holm-Sidak multiple comparisons tests, significance was indicated by a P value less than a critical value that is adjusted downward from 0.05 based on the rank of the value and the number of comparisons made.

	Results

Top Abstract Materials and Methods Results Discussion References

 
Individual FMO3 SNP Frequencies. Extending the previously reported analysis of common FMO3 upstream variant haplotypes (Koukouritaki et al., 2005), complete genotyping results for 17 FMO3 SNPs in three study populations are shown in Table 1 (the g.-2099A>G SNP was not observed in any of the studied populations and as such, it is not included in Table 1). None of the SNPs deviated from the proportions predicted by Hardy-Weinberg equilibrium. As was observed for the upstream SNPs, significant differences in the frequency of individual SNPs within the FMO3 structural gene were observed among the study groups. Of the two novel SNPs predicted to impact protein structure (Koukouritaki et al., 2005), the g.72G>T (E24D) variant was not observed in the African-American or Hispanic (Mexican descent) population groups and only occurred in the non-Latino white group at a frequency of 0.5%. In contrast, the N61K variant occurred at a frequency of 5.2% in the non-Latino white and 3.5% in the African-American population groups but was not observed in Hispanics (Mexican descent). The g.23613G>T (K416N) variant was only observed in the non-Latino white study group at an allelic frequency of 0.2%. The observed frequencies of the g.15167G>A (E158K), g.18281G>A (V257M), and g.21443A>G (E308G) variants were consistent with a previous report in which a haplotype analysis was restricted to these three SNPs (Cashman et al., 2001).

The synonymous variant K167K was not observed in any of the study populations, whereas the T108T synonymous variant was only observed in the non-Latino white group and only at 0.5%. Yet, the K167K and T108T SNPS were observed in four of 48 and two of 48 alleles tested in the Coriell Polymorphism Discovery Resource samples, respectively. This observation suggests these synonymous variants may be present at a significantly higher frequency in one of the other population groups represented in this resource (i.e., Asian Americans or Native Americans). Two other synonymous SNPs, i.e., g.15146C>T (S147S) and g.21375C>T (N285N), occurred at relatively high frequencies that differed among the population groups examined.

Based on an analysis of posterior estimates of population baseline recombination rates and calculated multiplicative factor parameters with posterior quantiles, no significant recombination was detected within the FMO3 locus beyond the background population rate at an level of 0.05 (Fig. 1). However, both the Hispanic and African-American samples indicated the possibility of a recombination "hot spot" between FMO3 position 15,437 and 18,281 and between both 15,437 and 18,281 and 18,281 and 21,375, respectively. In each of these instances, the 25% posterior quantile for _j between the pairwise SNPs was larger than 1, implying that the posterior probability of increased recombination is greater than 75% (Fig. 1).

View larger version (21K): [in this window] [in a new window]   Fig. 1. Identification of possible FMO3 recombination hot spots. Posterior estimates of population-specific baseline recombination rates () and multiplicative factor parameters (j) between adjacent SNPs were determined with 5% (), 10% (), and 25% () posterior quantiles (PQ) for the non-Latino white (A), Hispanic of Mexican descent (B), and African-American (C) populations (Li and Stephens, 2003). Numbers along the ordinate are indicative of the space between adjacent SNP pairs, beginning with g.-2650C>G and g.-2589C>T and proceeding in a 5'-to-3' direction along the FMO3 gene (see Table 1).

FMO3 Haplotype Analysis. Based on the determined allelic frequencies of the upstream and structural variants among the three populations included in this study, FMO3 haplotypes were inferred using PHASE, version 2.1 (Table 3). Five haplotypes were deduced in the Hispanic study population that occurred at a frequency greater than 5% and accounted for 90% of the variability within this population. Six haplotypes were inferred in the non-Latino white population that exhibited a frequency greater than 5% and accounted for 69% of the variability within this group. Finally, in the African-American study population, five haplotypes were deduced that occurred at a frequency equal to or greater than 5% and accounted for 63% of the population variation. AMOVA, which yields a statistic analogous to Wright's fixation index, indicated the percentage variance at the FMO3 locus among the populations studied was 0.0495% (bootstrap 95% confidence interval 0.02518, 0.07790), suggesting minimal difference in diversity. A calculation of FMO3 haplotype diversity was consistent with these results, but it did reveal the highest level of diversity in African Americans, followed by non-Latino whites and Hispanics of Mexican descent (Table 3). The 95% asymptotic confidence intervals calculated to provide an indication of the variance in haplotype diversity within each population should be interpreted with some caution, however, because they neglect the inherent variance in estimating unknown phase haplotypes.

Significant differences in haplotype frequencies were observed among the three study populations. Of particular interest were the differences observed in haplotypes involving SNPs known to have a functional impact on either gene expression or enzyme activity. The H2A and H2B haplotypes, consisting of the high-activity promoter variant (g.-2650C>G, g.-2543T>A, and g.-2177G>C) (Koukouritaki et al., 2005) and either two synonymous variants, S147S and N285N (H2A), or a single synonymous variant, S147S (H2B), occurred at a substantially higher frequency in the Hispanic (Mexican descent) population (26.6%) versus either the non-Latino white or African-American groups (4.6 and 4.5%, respectively).

The E158K and E308G variants are reported to have altered kinetic constants for the oxidation of several substrates consistent with a 40 to 70% loss of activity. Moreover, the compound variant has been reported to have an even more substantial 2- to 4-fold loss of enzyme activity (for review, see Koukouritaki and Hines, 2005). The inferred H3 haplotype frequency, consisting of the single SNP, g.+15,167G>A encoding the E158K variant, was highest in the Hispanic population (22.5%) but was common in all three population groups (Table 3). In contrast, the FMO3 H5B haplotype that includes both the E158K and E308G variants exhibited inferred frequencies of 5.4 and 7.9% in the Hispanic and non-Latino white population groups, respectively, but less than 5% in the African-American group. Yet, the FMO3 H6 haplotype, consisting of the E158K and the N285N variants, exhibited an inferred frequency of 14.6% in the African-American study population but less than 5% in Hispanics and was not inferred to be present in non-Latino whites. Interestingly, the sums of the deduced population frequencies for H3, H5B, and H6 that were 5% are similar; 27.9% in Hispanics of Mexican descent, 22.7% in non-Latino whites, and 23.7% in African Americans.

The previously described FMO3 promoter variants (Koukouritaki et al., 2005) may mask or exacerbate the effect of the structural variants. The high-activity promoter variant (previously reported as haplotype 2; -2650C>G, -2543T>A, -2177G>C) was inferred in eight haplotypes in the Hispanic (Mexican descent) population and 14 haplotypes in the nonLatino white population with a total frequency of 29.0 and 6.7%, respectively. Two of these haplotypes in the Hispanic group and four in the non-Latino white group also contain at least one of the reduced activity structural variants (E158K and/or E308G), but they occur at an inferred frequency of 1% or less in both groups. In contrast, the high-activity promoter variant was inferred in 21 African-American haplotypes with a total frequency of 11.7%. Importantly, 13 of these haplotypes also contain at least one of the reduced activity structural variants (E158K and/or E308G) and had a total inferred frequency of 5.6%.

Two low-activity promoter haplotype variants were previously identified, referred to as haplotype 8 (-2589C>T and -2106G>A) and 15 (-2106G>A and -1961T>C), both of which occurred at inferred frequencies <5% (Koukouritaki et al., 2005). Not surprisingly, neither one of these haplotypes were inferred to occur at frequencies >5% in the current study. However, given their predicted impact on FMO3 expression, a more important measure of their significance is the total frequency of these SNP combinations in all inferred haplotypes. In the Hispanic (Mexican descent) and non-Latino white population groups, only the -2589C>T, -2106G>A SNP combination was inferred in a total of four and six haplotypes and at frequencies of <1 and 4%, respectively. Both low-activity promoter variants were inferred in the African-American population group; the -2589C>T, -2106G>A in three haplotypes (total frequency <1%) and the -2106G>A, -1961T>C in seven haplotypes (total frequency of 4%). Three of the latter haplotypes also contained one of the reduced activity structural variants, i.e., E158K, E308G, or N61K.

Neither the FMO3 E24D nor N61K variants were inferred in any of the common FMO3 haplotypes. However, given the near complete loss of catalytic activity observed with the N61K variant, the individual SNP frequency for this variant is the most important measure of its functional impact. Although absent in the Hispanic population study group, the N61K variant was inferred in 25 non-Latino white haplotypes, total frequency of 5.2%, 12 of which (total frequency of 3.0%) also involve one or more of the FMO3 reduced activity variants, E158K and/or E308G. Likewise, the N61K variant was inferred in 17 haplotypes in the African American population group at a total frequency of 3.5%. Nine of these 17 haplotypes also contain one or more of the E158K and/or E308G structural variants at an inferred frequency of 2.2%.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
Genetic variation in drug-metabolizing enzymes is a well known determinant of interindividual differences in drug and toxicant disposition and response. In many cases, the differences are substantial with important clinical implications (Evans and McLeod, 2003). Underlying knowledge of such genetic variation and its associated functional changes is increasingly considered critical for drug discovery and development (Roses, 2004) and for the design of effective therapeutic regimens (Hisamuddin et al., 2005). Few previous studies have attempted to examine FMO3 haplotype structure and its impact on overall gene function. The results of an FMO3 SNP discovery effort that included all exons, splice sites, and approximately 1 kilobase pair of upstream information was reported recently (Koukouritaki et al., 2005). This effort also examined the potential functional significance of upstream haplotypes identified in three study populations. The current study extended these findings, testing the functional impact of two novel structural variants predicted to effect secondary structure, but perhaps more importantly, integrating the previous and current data into an overall FMO3 haplotype structure that allows a better prediction of genotype/phenotype relationships.

Using expressed recombinant FMO3, the kinetic parameters of the FMO3 E24D, N61K, and K416N variants were determined using four different substrates (methimazole, ethylenethiourea, trimethylamine, and sulindac). The E24D variant had a modest effect on FMO3 enzyme activity with an increase in K_cat and resultant increase in catalytic efficiency with most substrates. The results observed with the FMO3 K416N variant were much more equivocal and were substrate-dependent. In contrast, the N61K variant exhibited a complete loss of catalytic activity toward both trimethylamine and ethylenethiorurea and a 30- to 40-fold decrease in catalytic efficiency with both methimazole and sulindac. These observations are consistent with the previous report by Dolphin et al. (2000) who reported that substitution of a serine for asparagine at the same position (N61S) abolished the capacity of FMO3 to catalyze the N-oxidation of trimethylamine. Thus, similar to the FMO3 N61S variant, the N61K variant is predicted to contribute to the incidence of trimethylaminuria, an autosomal recessive disorder that is characterized by a deficiency in FMO3 enzyme activity and an inability of individuals to metabolize malodorous trimethylamine to the odorless trimethylamine N-oxide (Ayesh et al., 1993). Predictably, the N61K variant was not inferred in any of the common FMO3 haplotypes. However, the individual SNP frequency for this variant is of greater interest, given its functional impact. Not observed in the Hispanic population, the N61K variant occurred at an allelic frequency of 5.2 and 3.5% and was inferred in 25 and 17 haplotypes in the non-Latino white and African-American populations, respectively. Interestingly, 12 of the non-Latino white and nine of the African-American haplotypes also involved at least one of the previously characterized hypomorphic structural variants, g.15167G>A (E158K) and g.21443A>G (E308G), which would, if anything, exacerbate the loss of activity. In such individuals, genotype/phenotype correlations based on the presence or absence of the E158K and/or E308G variants alone also would overestimated the contribution of these latter SNPs to the phenotype.

Eswaramoorthy et al. (2006) recently reported the crystal structure of Schizosaccharomyces pombe FMO as enzyme-FAD, enzyme-FAD-NADPH, and enzyme-FAD-methimazole complexes. Alignment of human FMO3 with the S. pombe FMO amino acid sequence revealed an overall identity of only 21%, but conservation of the asparagine residue wherein the human FMO3 N61 position is analogous to yeast N91. Indeed, an alignment of 56 reported eukaryotic FMO sequences from 19 species, including Schizosaccharomyces pombe (Eswaramoorthy et al., 2006), Saccharomyces cerivisiae (AY357358 [GenBank] ), Caenorhabditis elegans (AJ582070 [GenBank] , AJ582071 [GenBank] , AJ582072 [GenBank] , AJ581300 [GenBank] , AJ582073 [GenBank] , NM_069567 [GenBank] , NM_069571 [GenBank] , NM_066955 [GenBank] , NM_073969 [GenBank] , and NM_070951 [GenBank] ), Drosophila melanogaster (NM_138015 [GenBank] and NM_136373 [GenBank] ), Crassostrea gigas (AJ585074 [GenBank] ), Danio rerio (NM_198910 [GenBank] ), Xenopus tropicalis (NM_001030424), Galus galus (AJ431490 [GenBank] ), Canis familiaris (AF384053 [GenBank] , XM_537197 [GenBank] , and AF384054 [GenBank] ), Mus musculus (NM_010231 [GenBank] , NM_018881 [GenBank] , NM_008030 [GenBank] , NM_144878 [GenBank] , and NM_010232 [GenBank] ), Rattus norvegicus (NM_012792 [GenBank] , NM_144737 [GenBank] , NM_053433 [GenBank] , NM_144561 [GenBank] , and NM_144739 [GenBank] ), Oryctolagus cuniculus (M32030 [GenBank] , M32029 [GenBank] , L10391 [GenBank] , L10392 [GenBank] , and L08449 [GenBank] ), Cavia porcellus (L10037 [GenBank] and L37081 [GenBank] ), Sus scrofa (NM_214064 [GenBank] ), Bos taurus (XM_866868 [GenBank] , XM_580516, NM_174057 [GenBank] , XM_586631 [GenBank] , and XM_598393 [GenBank] ), Macaca mulatta (U59453 [GenBank] and AY063498 [GenBank] ), Pongo pygmaeus (CR860815 [GenBank] and CR857503 [GenBank] ), Pan troglodytes (XM_513039 [GenBank] and NM_001009008 and NM_001009092), and Homo sapiens (NM_002021 [GenBank] , NM_001460 [GenBank] , NM_006894 [GenBank] , NM_002022 [GenBank] , and NM_001461 [GenBank] ) revealed 100% conservation of this asparagine, consistent with the data from the S. pombe FMO crystal structures that this asparagine moiety is the only amino acid residue directly involved in catalysis (Eswaramoorthy et al., 2006). The glutamate at human FMO3 position 24 is not as highly conserved as asparagine position 61, being observed in the FMO from all of the above-mentioned vertebrate species, but not in most of the nonvertebrate FMOs. Furthermore, the impact of the E24D variant on FMO3 activity was less than predicted in a previous report (Koukouritaki et al., 2005). Human FMO3 glutamate position 24 aligns with S. pombe glutamate position 28, which is involved in a turn between helix 1 and sheet 2, but structurally, is positioned well away from both the cofactor and substrate binding sites (Eswaramoorthy et al., 2006). Thus, its modest but consistent effect of increasing the FMO3 K_cat is likely due to minor confirmation changes in the overall protein structure that facilitate catalysis. Somewhat similarly, the lysine at position 416 in human FMO3 also is not that highly conserved among the 56 FMO proteins examined, although it is 100% conserved among mammalian FMO3 enzymes. Aligning the human FMO3 and S. pombe FMO primary sequence suggests that FMO3 lysine position 416 would fall within the carboxyl-terminal end of helix 6 (Eswaramoorthy et al., 2006), and, given the lack of predicted impact on secondary structure (Koukouritaki et al., 2005), it is not surprising that this amino acid substitution has minimal effects on FMO3 catalytic properties. Although the comparisons between human FMO3 and S. pombe FMO are interesting and consistent with the reported functional data, because the overall sequence identity is only 21%, conclusions drawn from these comparisons should be taken cautiously.

An important objective of the current study was to determine whether linkage between previously identified hypo- or hypermorphic promoter variants (Koukouritaki et al., 2005) and any of the FMO3 structural variants might contribute to discrepancies between in vitro and in vivo observations that were based on the more common analysis of structural variants alone. In the African-American population, the high-activity promoter haplotype (previously reported as haplotype 2; Koukouritaki et al., 2005) was inferred to occur in 21 haplotypes with a total frequency of 11.7%. Importantly, 13 of these haplotypes also contain at least one of the reduced activity structural variants (E158K and/or E308G) and occurred at a total frequency of 5.6%. In such individuals, genotype/phenotype association based solely on the analysis of structural variants would be misleading as one would predict the high-activity promoter variant would mask the impact of the structural variants. Similar linkage in the Hispanic and non-Latino white population groups was much less frequent, occurring at less than 1%.

Structural variants previously characterized as causative for reduced activity phenotypes based on in vitro studies would be exacerbated by low-activity promoter variants. Indeed, in the non-Latino white population, the promoter SNP cluster exhibiting low activity (previously reported as haplotype 8) occurred in six haplotypes with a total frequency of 4%, five of which contained at least one of the reduced activity structural variants (E158K and/or E308G). In the African-American population, the promoter SNP cluster exhibiting low activity previously reported as haplotype 15 occurred in seven haplotypes with a total frequency of 4%. Two of these haplotypes also contained at least one of the reduced activity structural variants (E158K and/or E308G), whereas one haplotype contained the N61K variant. The absence of low activity promoter variants in the Hispanic (Mexican) study population combined with the absence of the N61K allele, as well as the relative abundance of the high-activity haplotype allele, would be consistent with higher mean FMO3 expression in the Hispanic (Mexican) population.

FMO3 haplotype diversity was highest in African Americans, followed by non-Latino whites and Hispanics of Mexican descent. The greater genetic diversity observed among the African Americans is consistent with the findings of other studies (Salisbury et al., 2003). A possible recombination hot spot also was identified with a probability of a recombination rate above the African-American and Hispanic population baselines of at least 75%. Although these hot spots are not statistically significant at the 95% level, they merit consideration given the fact that the current study was not powered to detect small increases in recombination (Li and Stephens, 2003).

In summary, the reported haplotype analysis provides strong evidence that genetic variation in both the FMO3 promoter and structural gene contributes to the observed interindividual differences in FMO3 expression. Furthermore, the haplotype analysis strongly suggests that variation in promoter sequences can modulate the effects of previously characterized structural variants and should be taken into account when assessing genotype/phenotype association studies, or even when prescribing therapeutics for which FMO3 is important for disposition, e.g., sulindac (Hamman et al., 2000), itopride (Mushiroda et al., 2000), or pyrazoloacridine (Reid et al., 2004). Although there are minimal differences in overall FMO3 diversity among the population groups studied, significant differences in individual haplotype frequencies were observed that would contribute to interpopulational differences in the FMO3-dependent metabolism of therapeutics, environmental chemicals, and dietary constituents.

	Footnotes

 
This study was supported in part by Public Health Service Grants CA53106 (to R.N.H.) and HL38650 (to D.E.W.) and with funds from the Children's Research Institute, Children Hospital and Health Systems.

Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.

doi:10.1124/jpet.106.112268.

ABBREVIATIONS: FMO, flavin-containing monooxygenase; SNP, single-nucleotide polymorphism; SBE, single base extension; Sf, Spodoptera frugiperda; AMOVA, analysis of molecular variance; ANOVA, analysis of variance; K_cat, catalytic rate constant; K_m, Michaelis-Menten constant.

The online version of this article (available at http://jpet.aspetjournals.org) contains supplemental material.

Address correspondence to: Dr. Ronald N. Hines, Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226. E-mail: rhines{at}mail.mcw.edu

	References

Top Abstract Materials and Methods Results Discussion References

 

Ayesh R, Mitchell SC, Zhang A, and Smith RL (1993) The fish-odour syndrome: biochemical, familial and clinical aspects. Br Med J 307: 655–657.[Medline]

Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254.[CrossRef][Medline]

Cashman JR, Zhang J, Leushner J, and Braun A (2001) Population distribution of human flavin-containing monooxygenase form 3: gene polymorphisms. Drug Metab Dispos 29: 1629–1637.[Abstract/Free Full Text]

Cherrington NJ, Cao Y, Cherrington JW, Rose RL, and Hodgson E (1998) Physiological factors affecting protein expression of flavin-containing monooxygenases 1,3, and 5. Xenobiotica 28: 673–682.[CrossRef][Medline]

Dixit A and Roche TE (1984) Spectrophotometric assay of the flavin-containing monooxygenase and changes in its activity in female mouse liver with nutritional and diurnal conditions. Arch Biochem Biophys 233: 50–63.[CrossRef][Medline]

Dolphin CT, Janmohamed A, Smith RL, Shephard EA, and Phillips IR (2000) Compound heterozygosity for missense mutations in the flavin-containing monooxygenase 3 (FMO3) gene in patients with fish-odour syndrome. Pharmacogenetics 10: 799–807.[CrossRef][Medline]

Eswaramoorthy S, Bonanno JB, Burley SK, and Swaminathan S (2006) Mechanism of action of a flavin-containing monooxygenase. Proc Natl Acad Sci USA 103: 9832–9837.[Abstract/Free Full Text]

Evans WE and McLeod HL (2003) Drug therapy: pharmacogenomics–drug disposition, drug targets, and side effects. N Engl J Med 348: 538–549.[Free Full Text]

Excoffier L, Laval G, and Schneider S (2005) Arlequin version 3.0: An integrated software package for population genetics data analysis. Evol Bioinform Online 1: 47–50.

Geweke J (1992) Evaluating the accuracy of sampling-based approaches to calculating posterior moments, in Bayesian Statistics 4: Proceedings of the Fourth Valencia International Meeting (Bernardo JM, Berger JO, Dawid AP, and Smith AFM eds), pp 169–194, Oxford University Press, Oxford, UK.

Hamman MA, Haehner-Daniels BD, Wrighton SA, Rettie AE, and Hall SD (2000) Stereoselective sulfoxidation of sulindac sulfide by flavin-containing monooxygenases. Biochem Pharmacol 60: 7–17.[CrossRef][Medline]

Henderson MC, Krueger SK, Siddens LK, Stevens JF, and Williams DE (2004) S-Oxygenation of the thioether organophosphate insecticides phorate and disulfoton by human lung flavin-containing monooxygenase 2. Biochem Pharmacol 68: 959–967.[CrossRef][Medline]

Hernandez D, Janmohamed A, Chandan P, Phillips IR, and Shephard EA (2004) Organization and evolution of the flavin-containing monooxygenase genes of human and mouse: identification of novel gene and pseudogene clusters. Pharmacogenetics 14: 117–130.[CrossRef][Medline]

Hines RN (2006) Developmental and tissue-specific expression of human flavin-containing monooxygenase 1 and 3. Expert Opin Drug Metab Toxicol 2: 41–49.[CrossRef][Medline]

Hisamuddin IM, Wehbi MA, Schmotzer B, Easley KA, Hylind LM, Giardiello FM, and Yang VW (2005) Genetic polymorphisms of flavin monooxygenase 3 in sulindac-induced regression of colorectal adenomas in familial adenomatous polyposis. Cancer Epidemiol Biomarkers Prev 14: 2366–2369.[Abstract/Free Full Text]

Hukkanen J, Dempsey D, Jacob P, 3rd, and Benowitz NL (2005) Effect of pregnancy on a measure of FMO3 activity. Br J Clin Pharmacol 60: 224–226.[CrossRef][Medline]

Koukouritaki SB and Hines RN (2005) Flavin-containing monooxygenase genetic polymorphism: impact on chemical metabolism and drug development. Pharmacogenomics 6: 807–822.[CrossRef][Medline]

Koukouritaki SB, Poch MT, Cabacungan ET, McCarver DG, and Hines RN (2005) Discovery of novel flavin-containing monooxygenase 3 (FMO3) single nucleotide polymorphisms and functional analysis of upstream haplotype variants. Mol Pharmacol 68: 383–392.[Abstract/Free Full Text]

Koukouritaki SB, Simpson P, Yeung CK, Rettie AE, and Hines RN (2002) Human hepatic flavin-containing monooxygenase 1 (FMO1) and 3 (FMO3) developmental expression. Pediatric Res 51: 236–243.[Medline]

Krueger SK and Williams DE (2005) Mammalian flavin-containing monooxygenases: structure/function, genetic polymorphisms and role in drug metabolism. Pharmacol Ther 106: 357–387.[CrossRef][Medline]

Krueger SK, Yueh M-F, Martin SR, Pereira CB, and Williams DE (2001) Characterization of expressed full-length and truncated FMO2 from Rhesus monkey. Drug Metab Dispos 29: 693–700.[Abstract/Free Full Text]

Li N and Stephens M (2003) Modeling linkage disequilibrium and identifying recombination hotspots using single-nucleotide polymorphism data. Genetics 165: 2213–2233.[Abstract/Free Full Text]

Mushiroda T, Douya R, Takahara E, and Nagata O (2000) The involvement of flavin-containing monooxygenase but not CYP3A4 in metabolism of itopride hydrochloride, a gastroprokinetic agent: Comparison with cisapride and Mosapride citrate. Drug Metab Dispos 28: 1231–1237.[Abstract/Free Full Text]

Nei M (1978) Estimation of average heterozygoxity and genetic distance from a small number of individuals. Genetics 89: 583–590.[Abstract/Free Full Text]

Overby LH, Carver GC, and Philpot RM (1997) Quantitation and kinetic properties of hepatic microsomal and recombinant flavin-containing monooxygenases 3 and 5 from humans. Chem Biol Interact 106: 29–45.[CrossRef][Medline]

Park C-S, Kang J-H, Chung W-G, Yi H-G, Pie J-E, Park D-K, Hines RN, McCarver DG, and Cha Y-N (2002) Ethnic differences in allelic frequency for two flavin-containing monooxygenase 3 (FMO3) polymorphisms: Linkage and effects on in vivo and in vitro FMO activities. Pharmacogenetics 12: 77–80.[CrossRef][Medline]

Reid JM, Walker DL, Miller JK, Benson LM, Tomlinson AJ, Naylor S, Blajeski AL, LoRusso PM, and Ames MM (2004) The metabolism of pyrazoloacridine (NSC 366140) by cytochromes P450 and flavin monooxygenase in human liver microsomes. Clin Cancer Res 10: 1471–1480.[Abstract/Free Full Text]

Roses AD (2004) Pharmacogenetics and drug development: the path to safer and more effective drugs. Nat Rev Genet 5: 645–656.[CrossRef][Medline]

Salisbury BA, Pungliya M, Choi JY, Jiang R, Sun XJ, and Stephens JC (2003) SNP and haplotype variation in the human genome. Mutat Res 526: 53–61.[Medline]

Shimada T, Yamazaki H, Mimura M, Inui Y, and Guengerich FP (1994) Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians. J Pharmacol Exp Ther 270: 414–423.[Abstract/Free Full Text]

Stephens M and Donnelly P (2003) A comparison of bayesian methods for haplotype reconstruction from population genotype data. Am J Hum Genet 73: 1162–1169.[CrossRef][Medline]

Stephens M, Smith NJ, and Donnelly P (2001) A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 68: 978–989.[CrossRef][Medline]

Störmer E, Roots I, and Brockmöller J (2000) Benzydamine N-oxidation as an index reaction reflecting FMO activity in human liver microsomes and impact of FMO3 polymorphisms on enzyme activity. Br J Clin Pharmacol 50: 553–561.[CrossRef][Medline]

Tijet N, Boutros PC, Moffat ID, Okey AB, Tuomisto J, and Pohjanvirta R (2006) Aryl hydrocarbon receptor regulates distinct dioxin-dependent and dioxin-independent gene batteries. Mol Pharmacol 69: 140–153.[Abstract/Free Full Text]

Yeung CK, Lang DH, Thummel KE, and Rettie AE (2000) Immunoquantitation of FMO1 in human liver, kidney, and intestine. Drug Metab Dispos 28: 1107–1111.[Abstract/Free Full Text]

Zhang AQ, Mitchell SC, and Smith RL (1996) Exacerbation of symptoms of fishodour syndrome during menstruation. Lancet 348: 1740–1741.[Medline]

Zschocke J, Kohlmüeller D, Quak E, Meissner T, Hoffmann GF, and Mayatepek E (1999) Mild trimethylaminuria caused by common variants in FMO3 gene. Lancet 354: 834–835.[Medline]

This Article Abstract Full Text (PDF) Data supplement All Versions of this Article: jpet.106.112268v1 320/1/266    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Koukouritaki, S. B. Articles by Hines, R. N. Search for Related Content PubMed PubMed Citation Articles by Koukouritaki, S. B. Articles by Hines, R. N.